Selection bias in observational studies of aerosol-cloud interactions


Satellite observations of aerosol-cloud interactions are hampered by the inability to detect aerosols in the presence of clouds and precipitation. Thus, there is a systematic bias regarding which samples of aerosol observations are selected (selection bias). Yet, satellites are the most important observation tool to estimate aerosol radiative forcing because of their global coverage. The objective of this project is to study this selection bias in estimating the interaction between aerosols and clouds from MODIS satellite data using novel causal discovery algorithms and compare it to high resolution global climate model simulations that don't suffer from selection bias. 


Ulrike Lohmann (supervisor), Gabi Hegerl and Jakob Runge (co-supervisors), Niall Robinson (MetOffice, non-academic advisor)


3 months at DLR, 3 months at University of Edinburgh, 3 months at Met Office

Enrolment in Doctoral degree

PhD in atmospheric and climate science at ETH Zurich

teaser image zurichwhite bg
dlr institute jena